sign in sign up
<<
Home , Cognitivism (ethics), Ethical naturalism

1 Naturalism

Charles R Pigden

What Naturalism Is In Philosophy not only the doctrines but the definitions of the doctrines are subject to dispute. Thus it is with . My definition, therefore, is personal but not, I hope, idiosyncratic. By naturalism I mean more or less what other philosophers have in mind — though the more and the less will vary from case to case. Naturalism, then, is a cognitivist doctrine (or family of doctrines). It states that moral judgements are , capable of truth and falsity. Moral judgements purport to tell it like it is. Naturalism is thus opposed to non-, to and prescriptivism, which represent moral judgements variously as exclamations, psychological prods and quasi-commands. It is also (in a weak sense) a realist doctrine; that is, it takes some moral judgements to be true. It is thus opposed to the error-theory of J. L. Mackie which agrees that moral judgements make claims that are either true or false, but denies that any of them are true. , for the naturalist, is not a fiction, a mistake or a myth but a body of knowledge, or at least, of information. Finally, naturalism is (in a loose sense) a reductive doctrine. Though there are moral truths (i.e. true propositions) there are no peculiarly moral facts or properties (no distinctively moral states of affairs) over and above the facts and properties that can be specified using non-moral terminology. The contrast here is with 'intuitionist' philosophers such as G.E Moore (1874-1958): 'If I am asked "What is ?", my answer is that good is good and that is the end of the matter. Or if I am asked "How is good to be defined?", my answer is that it cannot be defined and that is all I have to say about it' (Moore, 1903, p. 6). Moore does not mean he cannot tell you what things are good. (He thinks for example that friendship and the contemplation of beauty are good.) No, he is making a metaphysical or ontological point: the goodness of good things consists in their possessing the property of goodness, a basic feature of which cannot be further analyzed or explained. Naturalists disagree. For them, goodness can be further analyzed or explained; reduced to something else or identified with some other property. Indeed naturalists think that goodness as Moore conceives it, a unique and sui generis property, does not exist. (And the same goes, naturally, for badness, rightness and wrongness.) But there consensus ends. Naturalists differ as to what good, etc. are to be reduced to, and how this reduction is to be carried out. There are hedonistic naturalists who reduce facts about goodness to facts about and pain. (The goodness of friendship consists in its conduciveness to pleasure.) There are Aristotelian naturalists who prefer (alleged) facts about human nature and human flourishing. (Friendship is good in that it is somehow consonant with human needs or human nature.) There are even theological naturalists, who think the goodness of friendship amounts to its being sanctioned by God. Naturalists, in short, resort to all sorts of supposed facts — sociological, psychological, scientific, even metaphysical and theological — so long as they are not driven back on a realm of irreducibly moral facts or properties. 2

Since some of these facts are metaphysical or supernatural, rather than natural in any ordinary sense of the word (facts about the natural ) you may well wonder how such a disparate group of moral theories came to wear the same label. The answer is historical. According to G. E. Moore they all stand accused of the (of which, more later). He called this (supposed) fallacy naturalistic because it was more common among philosophers of a narrowly naturalistic stamp — those who wished to base morals on the kinds of facts that science could countenance. But these were only a subclass of those who — according to Moore — commit the fallacy. Nevertheless the name has stuck. What is the impetus behind naturalism — what makes it attractive as a theoretical option? This is a difficult question to deal with since naturalism is not so much a doctrine as a family of doctrines held together by some shared theses. But a rough answer is this. Naturalists combine a yen for moral truth, that is, a conviction that some things really are right and others wrong, with a distaste for Moore's non-natural qualities of goodness, badness etc. Often this distaste is due to the scientific outlook and the conviction that nothing exists beyond what science licenses us to suppose. But it can be due to religious convictions, e.g. the belief that springs from God and cannot be separated from what he wills. Either way, Moore's peculiarly moral properties are unwelcome, and a reduction must be sought which bases moral truths on the preferred metaphysic. 3

1. Moral judgements are propositions. (They Non-cognitivism in its various forms (emotivism are true or false.) and prescriptivism) is false.

2. Some moral judgements are true. Nihilism or the 'error-theory' (due to J.L (Morality not a fiction.) Mackie) is false.

3.No irreducible moral facts or properties. Intuitionism(Moore's doctrine) is false.

Figure 1

2. No-'ought'-from-'is', or the of If morality can be boiled down to truths of some other kind, ethics is not 'autonomous'. Truth in morals is determined by the goings-on in some other realm. However, when naturalists and anti-naturalists quarrel about the autonomy of ethics, this is not (usually) what they have in mind. By 'the autonomy of ethics' they mean some such thesis as that 'ought' cannot be derived from 'is', or more generally that moral conclusions cannot be derived from non-moral premises, values from facts. This is often mixed up with the idea that moral words do not mean the same as non-moral words; that they lack 'naturalistic' synonyms. The issue is thus about inference rather than reduction; about what can be inferred from what. However, it is often assumed that if moral judgements can be derived from non-moral propositions, naturalism is true. If not, naturalism is false. This last claim is a mistake as I hope to show. Anti-naturalists take as their text a famous passage from Hume (1711-1776). Moralists (complains Hume) 'proceed for some time in the ordinary way of reasoning' with proofs 'of the being of God' or 'observations concerning human affairs', 'when of a sudden, I am surprised to find that instead of the usual copulations of propositions is and is not, I meet with no that is not connected with an ought, or an ought not ... as this ought or ought not expresses some new relation or affirmation 'tis necessary that it be observed and explained; and at the same time a reason should be given for what seems altogether inconceivable, how this new relation can be a deduction from others which are entirely different from it' (Hume, 1738, Book III, Part 1). This passage is often summed up in the slogan 'no-"ought"-from-"is"' and deified as 'Hume's Law'. It is alleged to do wonders — to point up a fundamental distinction between facts and values, to prove non- cognitivism and (above all) to refute naturalism. (This last is an odd claim, since Hume himself was a naturalist.) In fact it can do none of these things. For Hume is making a simple logical point. A conclusion containing 'ought' cannot (as a matter of logic) be derived from 'ought'-free premises. (The same, of course, goes for the other moral words.) Logic is conservative; the conclusions of a valid inference are contained 4 within the premises. You don't get out what you haven't put in. Hence if 'ought' appears in the conclusion of an argument but not in the premises, the inference is not logically valid.

Some anti-naturalists (notably R. M. Hare) have taken this 'is'/'ought' gap as a datum and sought to explain it by means of non-cognitivism. The reason you can't derive an 'ought' from an 'is', or, more generally, a moral conclusion from non-moral premises, is because moral judgements differ fundamentally from factual propositions. They don't (primarily) describe how the world is but prescribe how it should be — they are, in short, akin to orders. But we need not resort to non-cognitivism to explain this logical gulf. For there is a similar gap between conclusions about hedgehogs and premises which make no mention of them. You can't get 'hedgehog' conclusions from hedgehog-free premises (at least, not by logic alone). Yet nobody proposes a fact/hedgehog distinction or alleges that propositions about hedgehogs are not really propositions but a quaint subclass of commands. In both cases it is the conservativeness of logic that creates the gap, not any arcane difference in semantic kind. There is thus no more reason to convert moral judgements into quasi-commands than to subject hedgehog- propositions to the same fate. Ethics may be logically autonomous, but it shares this trait with every other kind of talk. But even this innocuous thesis has been challenged. A. N. Prior (1914-1969) proposed a number of logically valid inferences in which 'ought' does appear in the conclusion but not in the premises. Here is one:

(1) Tea-drinking is common in England. Therefore (2) Either tea-drinking is common in England or all New Zealanders ought to be shot. (Prior, 1976)

There is something very fishy about this inference. You can't help feeling that the conclusion is not genuinely concerned with 'ought' at all. After all, the following is also a consequence of (1):

(2') Either tea-drinking is common in England or all New Zealanders are hedgehogs.

Here we have apparently hedgehog-conclusions from a hedgehog-free premise. What this shows is that Prior's counter-examples do not just threaten the autonomy of ethics (no 'ought' from 'is') but the conservativeness of logic, the idea that in logic you don't get out what you haven't put in. But it also becomes apparent that the (italicized) last clauses in (2) and (2') are vacuous in a certain sense. Given the premises they can be filled in anyhow without making the conclusions 5 untrue or affecting the validity of the inference. We are now in a position to restate both the conservativeness of logic and the autonomy of ethics. First we define inference-relative vacuity. An expression occurs vacuously in the conclusion of a valid inference if it can be (uniformly) switched for any other expression of the same grammatical type without prejudice to the validity of the inference. (The italicized last clauses of (2) and (2') are vacuous in this sense, since in an inference from (1) each can be switched with the other.) The conservativeness of logic becomes the claim that no (non-logical) expression can occur non- vacuously in the conclusion of a valid inference unless it appears in the premises. This claim is susceptible of proof. The autonomy of ethics is simply the moral incarnation of this claim — no non-vacuous 'ought' from 'is'. (See Pigden, 1989.) Where does all this leave naturalism? In fine fettle. We have already seen that logical autonomy affords no support for non-cognitivism. Nor (despite a widespread philosophical belief) does it endanger naturalism. As we have seen, hedgehog-conclusions cannot be derived non- vacuously from hedgehog-free premises. But this does not entail that there is a realm of irreducible hedgehog-facts. (Hedgehogs are composite creatures whose workings can be explained in terms of their parts.) So why should we posit a realm of irreducible moral facts to explain why 'ought' cannot be derived from 'is'? The mistake is due to a confusion. For there are three forms of the autonomy of ethics, logical, semantic and ontological. These are often confounded. But only the last is incompatible with naturalism. Logical autonomy I have dealt with already. Semantic autonomy (semantic because concerned with meanings) is the thesis that moral words do not mean the same as any others, and furthermore that they cannot be paraphrased in a naturalistic (or non-moral) idiom. Ontological autonomy is the thesis that moral judgements, to be true, must answer to a realm of sui generis moral facts and properties. (I call it ontological autonomy because it is concerned with the kinds of things there must be.) Ontological autonomy is thus the reverse of naturalism, which insists that nothing so queer as Moore's moral properties is required to sustain the truths of morals.

Now logical autonomy is correct. But it is no threat to naturalism unless it entails ontological autonomy. The hedgehog parallel proves that it does not. Does logical autonomy entail semantic autonomy? Again no. Logic alone does not license us to infer conclusions about hedgehogs from premises dealing with Erinaceus europeus. Hedgehog-talk is logically autonomous. But this says nothing one way or the other about whether 'hedgehog' and ‘erinaceus europeus' are synonymous (as in fact they are). So too in morals. Conclusions about good cannot be logically derived from premises about pleasure. But this says nothing one way or the other about whether 'good' and 'pleasant' are synonymous (as in fact they are not).

'But this can't be right’, you may say. 'After all, if "good" did mean "pleasant" we could logically derive conclusions about good from premises about pleasure so long as these included the appropriate definition. Ethics would not be logically autonomous after all. So if it is logically autonomous, it follows that goodness cannot be defined. Which amounts to semantic autonomy.' (See Prior, 1949, p. 24.) This is a 6 mistake. For a definition, although verbal, and hence, in a sense, insubstantial, still constitutes an extra premise. So if 'good' appears in a conclusion as the result of a definition, this in no way violates logical autonomy. For it had to appear in at least one premise to get there. So too with 'hedgehog'. If hedgehog- conclusions are derived from premises about erinaceus europeus with the aid of a definition (in this case a true one), this in no way violates the logical autonomy of hedgehog-talk. For it had to appear in at least one premise to get there. The truth or otherwise of the definition does not affect the logic of the matter. Now, logical autonomy is about validity while semantic autonomy is about meanings. It is a doctrine concerned with definitions. And on the truth or otherwise of definitions, logic is not competent to decide. (See Pigden, 1989.)

But is semantic autonomy true? And if so, does this entail ontological autonomy and hence the falsehood of naturalism? This brings us back to —

3 The naturalistic fallacy In his famous Principia Ethica, G. E. Moore contended that most moralists have been naturalists and that all naturalists are guilty of a common fallacy. They have confused the property of goodness with the things that possess that property or with some other property that good things possess. This is what the naturalistic fallacy is; a mixing up of two distinct items. Moore has two main arguments for this.

(1) Suppose goodness were identical with some other property such as pleasantness. (Moore's favoured candidate is what we desire to desire, but we shall stay with pleasantness for shortness' sake.) Then 'good' and 'pleasant' would be synonymous. This would be known to every competent speaker. Hence the question 'Is what is pleasant, good?' would be unintelligible. (Here Moore's choice of words is unfortunate. It is not that the question would be senseless. Rather, to ask it would show a want of sense, or at any rate, of understanding. For the answer would be very obvious — yes.) The question would be a mere tautology with a question-mark tacked on the end. It would have the same meaning as 'Is what is pleasant, pleasant?' or 'Is what is good, good?' Now, Moore takes it to be obvious that the question 'Is what is pleasant, good?' is an open question; a question that it does make sense to ask. Hence 'good' does not mean 'pleasant'. Therefore goodness and pleasantness are distinct. And the same goes for the other properties with which goodness is identified.

A similar argument (dating back to 's Euthyphro) has been deployed against the theological 7 naturalist. Does 'X is right' mean 'X is commanded by God'? Then 'What God commands is right' means 'What God commands is what God commands', and the moral praise of God degenerates into a fanfare of tautologies. But 'What God commands is right' is not a tautology. Hence the naturalistic definition is false.

(2) If 'good' meant 'pleasant' then to say 'What is pleasant is good' would provide us with no extra reason for promoting pleasurable states of affairs. (It would amount to the tautology that what is pleasant is pleasant, and a tautology cannot provide a motive for action.) But to call pleasure good is to suggest some extra reason for promoting it. Therefore 'good' doesn't mean 'pleasant'. Again the argument can be generalized.

These arguments fail in their intended purpose. They do not establish ontological autonomy and hence the falsehood of naturalism. At best they establish semantic autonomy and thus disprove semantic naturalism — the thesis that moral facts can be reduced to non-moral facts, because moral words are synonymous with (combinations of) non-moral words. And there is some doubt even of this. Argument (1) assumes that if goodness is identical with pleasantness, 'good' and 'pleasant' must be synonymous. This is false. Water and H2O are identical. Yet 'water' is not synonymous with `H2O' even though they stand for the same stuff. 'Water' expresses a pre-scientific concept accessible to children and savages — roughly, the colourless, tasteless fluid that falls from the sky and is found in lakes and rivers. `H2O', by contrast, expresses a scientific notion. You can't understand it fully without a modicum of chemistry. People did not find out that water was H2O by meditating upon meanings. Empirical enquiry did the trick. So too with goodness and pleasantness. 'Good' may not be synonymous with 'pleasant'. But they might both stand for the same property. Semantic naturalism might be false but synthetic naturalism might be true. That is, moral properties might be identified with natural properties by means of empirical research rather than conceptual analysis. Thus semantic autonomy, which says that moral words do not mean the same as any others, does not entail ontological autonomy — that moral properties are not identical with any others. But there is more. Argument (1) also assumes that if two words or phrases are synonymous, this is known to every competent speaker. And we can grant that if synonymy is given a strict interpretation, this is so. But our concepts are not transparent to us. The rules they obey, the presuppositions they embody, are not things we are always fully conscious of. Thus it may be possible to give an analysis of a word — a conceptual breakdown — which won't be obvious to competent speakers, although in some sense it expresses the word's meaning. (Otherwise the 'paradox of analysis' would preclude useful conceptual analyses.) So it may be possible to formulate an analysis, X, of 'good' such that 'Are X things good?' is an open question (i.e. one a competent speaker might sensibly ask) even though X articulated the 8 meaning of 'good'. This puts semantic autonomy into question.

Argument (2) likewise supports the conclusion that 'good' is without naturalistic (i.e. non-moral) synonyms or paraphrases. But this is semantic autonomy, not the ontological autonomy that controverts naturalism. Moreover, the argument is suspect. It trades on the idea that 'good' conveys some requirement on action. If someone thinks something good, this usually disposes them to pursue or promote it (in the appropriate circumstances). Fair enough. But perhaps an analysis can be given which spells out this requirement without recourse to non-natural properties. 'Pleasant' may be a poor choice for this role, but this is not to say that a better can't be found. (For example, 'good' could mean required where the requirements relate to some goal all rational beings could be expected to share.) Of course, if such an analysis, X, could be constructed, 'What is X, is good' would be a tautology or conceptual truth, bereft of motivating power. But this is irrelevant. For such a proposition would not be designed to commend X-ness, or to indicate that X-ness was somehow required. Rather the purpose would be to articulate the requirements wrapped up in the predicate `good'; to explain why it is that calling things 'good' usually suggests a reason to promote them.

To sum up. Logical autonomy (no-'ought'-from-'is') can be proved in an amended form. But this does not endanger naturalism. Semantic autonomy may be true too. Perhaps 'good' and the other moral words lack naturalistic or non-moral synonyms or paraphrases — at least, synonyms or paraphrases with which they are strictly synonymous. This disposes of semantic naturalism, but leaves other kinds of naturalism intact. So naturalism could be true, despite Hume and despite G. E. Moore.

Thus there is no need for naturalists to evade the arguments of Moore and Hume by making out that the moral 'ought' and the predicative 'good' (Hume's 'ought' and Moore's 'good') are senseless. (Something Anscombe in 'Modern moral philosophy' and Geach in '' try to do.) Insofar as they are valid, Hume's arguments, and Moore's too, are compatible with naturalism. Formal attempts to refute naturalism having failed, it remains a live option.

4. Variants of naturalism I conclude with a survey of the leading brands of naturalism. My purpose is exposition rather than critique, but I won't forswear critical comments.

(A) The best bet, given our objections to Moore, looks like synthetic naturalism. 'Good' (and the same goes for the other moral words) does not mean the same as any naturalistic 'X'. Nevertheless there is (or could be) some naturalistic predicate 'X' which

stands for the same property. (Just as 'water' and 'H20' don't express the same concept, though they stand for the same stuff.) The identity between goodness and X-ness would not be analytic, holding in of the meanings of words, but synthetic, a matter of (empirical?) fact. R. M. Adams, a theological naturalist, tries to rehabilitate the Divine 9

Command theory of ethics using this idea (Adams, 1981). To say 'X is right' does not mean that God commands it. Hence 'What God commands is right' is not tautologous. Nevertheless, rightness and being commanded by God come to much the same thing. We apprehend that certain actions are right and subsequently discover (through revelation or rational theology) that what we were aware of was the Voice of God. But the problems with this approach are best seen in a secular setting. If goodness is to be identified with a naturalistic something by means of empirical enquiry, 'good' must express an empirical concept. That is, 'good' must be defined in such a way that goodness impinges upon us and can figure within an empirical theory. For instance, it might be defined as the cause of certain effects, the 'goodness-phenomena'; which phenomena can (at least sometimes) be sensed by human beings. (Otherwise empirical enquiry could not determine whether goodness and X-ness applied to the same things.) But this, in turn, presupposes that 'good' can be subject to a naturalistic analysis; that in some sense it means the same as a paraphrase which relates it to sensory evidence. In other words, Moore must be wrong and 'good' must be analysable — although it need not be strictly synonymous with the paraphrase which constitutes the analysis. The best candidate is the kind of theory that (on some interpretations) was propounded by . 'Good' is not defined in terms of our ordinary sensations — sight, smell, hearing, touch — but in terms of internal sensations of approval and disapproval. 'Good' is (roughly) what every informed and impartial spectator would approve of (and hence what we do tend to approve of, when we rid ourselves of partial passions and try to work out what is right.) Of course, once we have got such an analysis, naturalism is already home and hosed. We don't need to go on to establish a synthetic identity between goodness, so analysed, and some other, natural property. If being good means being approved of by an spectator, we need not determine what the spectator would approve of to know we have moral truths without irreducible moral facts (though what the spectator approves of will be of considerable practical import). Morals boils down to an idealized psychology. Naturalism is vindicated.

The theory is beset by problems. Here are two. (a) It is not clear that all human observers however impartial and well-informed would agree in their reactions. Hence the possibility opens up that no (or very few) actions are good, since none would excite the approbation of every spectator. We revert to the error theory we were trying to avoid (Section i). (b) The analysis looks circular. For to approve of something is to think (or feel) it is good or right. Hence the analysis of 'good' contains the very concept it is designed to explicate.

(B) (This theory is suggested by the writings of G. E. M. Anscombe, though I am not sure whether she would agree with it. Sabina Lovibond, Realism and Imagination in Ethics, believes something of the sort.) Suppose I have ordered some potatoes and the grocer has 10

delivered them and sent me the bill. Then I owe the grocer money. Now, what makes it the case that I am in debt to the grocer? (i) A history of personal transactions between the grocer and me; (ii) some general rules, holding in virtue of social institutions and conventions, about how debts are contracted; and (iii) the absence of special conditions which might cancel or invalidate the debt. In other words, a complex of social facts which boil down to human dispositions and actions. There is nothing more to my being in debt than this; no special non- natural realm of debt-facts — just human actions in the context of human institutions; which institutions are themselves dependent on continued human action. The naturalist hypothesis, then, is that some moral judgements (It was wrong of you to tell that lie') are like 'I owe the grocer money' — true in virtue of individual human actions in the context of our institutions or language games. Other, more general, judgements ('It is usually wrong to lie') resemble the propositions stating the rules for contracting a debt — they hold true in virtue of human conventions and institutions, shared social practices. Thus we have moral truths, both general and particular, without peculiarly moral facts. Morals boils down to a sophisticated sociology. Naturalism is vindicated.

The problem here (as Anscombe seems to realize) is relativity. Different societies have different moral institutions sustaining different moral codes. Do we index moral truth to the codes of particular societies, so that there is rightness for the Azande, and rightness for the Aussies, but no such thing as rightness period? What about divided societies where different codes and institutions compete (like most Western societies)? And what becomes of moral dissidents? It seems that they are not merely bad (because they disregard the only moral truth there is) but incoherent, since we can give no content to the dissident's thought that the code her society sustains is not really right. (Sabina Lovibond wrestles with this problem for 100 pages. She does not come off the victor.) Finally, this approach has the unfortunate consequence that the moral opinions of Margaret Thatcher and the Ayatollah Khomeini do not really contradict one another, since both are indexed to the institutions of their own society. Suppose we remove the indices and just talk about right or wrong. Then in the absence of cross-cultural institutions determining a super-cultural ethic there is no such thing as moral truth. But a naturalistic basis for moral truth was the of the enterprise.

(C) Lastly I turn to a form of neo-Aristotelian naturalism hinted at by P. T. Geach and developed in a rather haphazard way by M. Midgley and others. Rather than worrying about what things are good and what goodness is (a mythical property anyway) we should be concerned about what it is to be a good human being and to perform good human acts. Geach seems to think that these notions carry no metaphysical baggage and can be given a purely natural explication. Man (presumably) has a function, and once we find out what that function is, we will have no trouble finding out what being a good man consists in. Geach's religious beliefs have led him away from this project 11

towards a of ethics. It has been left to others, principally Midgley, to develop these ideas in a secular and biological context. Drawing on ethological literature, she suggests that, given our natures, there are constraints on the kinds of lives humans will find fulfilling, and hence on human action. Morality (it seems) can be boiled down to a refined biology. Naturalism is vindicated.

I can only record my impression that after thirty years of effort, not a lot of progress has been made with this programme. The writings of Geach, Midgley and their allies are 'suggestive' but nothing more. First, it is far from obvious that 'good human' can be given a naturalistic reading without relapsing into the sociologistic naturalism discussed above. Secondly, I doubt whether the Aristotelian concept of a function can be revived within modern biology, at least for such a flexible and acculturated creature as man. And this needs to be done if the programme is to achieve anything concrete. If morals are to be based on biology we must somehow generate a set of requirements out of human nature. These must be (i) reasonably specific; (ii) rationally binding or at least highly persuasive; and (iii) morally credible. So far, no such biologically based requirements have been provided. We live in hope. This concludes my survey of the leading brands of naturalism. They are not so much wrong in as deficient in detail. As we have seen, there is no sweeping argument which condemns the naturalistic enterprise as unworkable. It can't be excluded on formal grounds. Nevertheless the variants currently on offer leave much to be desired. But this doesn't dispose of naturalism either, since there may be other options. So for all the apocalyptic pronouncements of Moore and the non-cognitivists, who dismissed naturalism (almost by definition) as fallacious, naturalism nowadays is very much a going concern. Whether it is right or wrong only time — and further argument — will tell. 12

References Adams, R. M.: 'Divine command metaethics as necessary a posteriori', Divine Commands and Morality, ed. P. Helm (Oxford: Oxford University Press, 1981). Anscombe, G. E. M.: 'On brute facts', 'Modern moral philosophy' and 'On the source of the authority of the state', in her Ethics, Religion and Politics (Collected Papers, Vol. 3), (Oxford: Basil Blackwell, 1981). Geach, P. T.: 'Good and evil', Theories of Ethics, ed. P. Foot (Oxford: Oxford University Press, 1967). Hare, R. M.: The Language of Morals (Oxford: Oxford University Press, 1952). Hume, D.: Treatise on Human Nature (1738); ed. L. A. Selby-Bigge (Oxford: Clarendon Press, 1978). Lovibond, S.: Realism and Imagination in Ethics (Oxford: Basil Blackwell, 1983). Mackie, J.: Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977). Midgley, M.: Beast and Man (Ithaca, NY: Cornell University Press, 1978). Moore, G. E.: Principia Ethica (Cambridge: Cambridge University Press, 1903). Pigden, C. R.: 'Logic and the autonomy of ethics', Australasian Journal of Philosophy, 67 (1989), 127-51. Plato: Euthyphro. Prior, A. N.: Logic and the Basis of Ethics (Oxford: Oxford University Press, 1949). Prior, A. N'The autonomy of ethics', in his Papers on Logic and Ethics, (London: Duckworth,(1976).

Further reading Carson, T. L.: The Status of Morality (Dordrecht: Reidel, 1984). Chisholm, R. M.: Brentano and Intrinsic Value (Cambridge: Cambridge University Press, 1986). Clark, S. R. L.: 'The lack of gap between fact and value', Proceedings of the Aristotelian Society, Supplementary Volume 54 (1980) 225-40. Foot, P.: and Vices (Oxford: Basil Blackwell, 1978). Geach, P. T.: The Virtues (Cambridge: Cambridge University Press, 1977). Hume, D.: Moral and , ed. D. Aiken (New York: Hafner, 1948). Lewis, D.: `Dispositional theories of value', Proceedings of the Aristotelian Society, Supplementary Volume 63 (1989), 113-137. Midgley, M.: 'On the lack of gap between fact and value', Proceedings of the Aristotelian Society, Supplementary Volume 54 (1980), 207-23. Pigden, C. R.: `Geach on "Good" ', Philosophical Quarterly, 40 (1990), 1-20. Putnam, H.: Reason, Truth and History (Cambridge: Cambridge University Press, (1981). See especially Chapter 9.